It is advantageous to control flow rate in chromatography instruments (e.g., high-performance liquid chromatography instrument or HPLC, ultra-performance liquid chromatography instrument or UPLC™, or supercritical fluid chromatography instrument or SFC) such that a loss of resolution or information is substantially minimized or avoided. For example, upstream of the point of analyte introduction or “sample injection”, in a “gradient mode” of liquid chromatography, solvent composition profiles are generated by pumps and corresponding controllers, where those composition profiles have a time-course and a corresponding volume-course. As the volume scale of chromatography decreases, the volume scale of the gradient profile likewise decreases. Thus in nanoscale chromatography, an entire solvent gradient profile may encompass only a few microliters of liquid volume. Within that few microliters is expected to reside at least a defined linear ramp of solvent composition, or perhaps a staircase of solvent composition steps. Assertion of that solvent composition profile, with good fidelity, onto the chromatography column, requires that the internal volumes of the intervening fluid paths are small relative to the volume of the gradient profile features, and extremely cleanly-swept. In order to improve results it is advantageous to avoid the presence of poorly-swept or “dead” volumes in these liquid systems. Emphasis is placed on the reduction or elimination of dead volumes throughout the chromatography system, to maximize system performance.
To control or monitor the flow through an instrument, it is sometimes desirable to include one or more sensors in communication with the fluid stream(s) used within the instrument. In a large-scale chemical process where one might seek to obtain, for example, a pressure measurement, it is typically straightforward to purchase a commercially-available sensor and to install the sensor into the system using readily available mounting flanges with screw-thread connectors. Many large-scale chemical processes are relatively immune to the introduction of incremental volumes (often milliliters) associated with such a measurement. The cleanliness with which that introduced volume is swept by the process stream may also not be particularly demanding (e.g., in a system where tens or hundreds of liters per minute of process flow exists, a transducer which has a few milliliters of internal volume, which is swept or exchanged once during the passage of a liter or more of the process stream, may be perfectly acceptable.)
One can readily appreciate that in a fluid chromatography instrument where an entire solvent gradient profile may encompass only a few microliters of volume, up to as much as a few milliliters of volume, such a transducer implementation is not acceptable.
Most commercially available pressure sensors capable of measurements up to tens or hundreds of megapascals (e.g., high-pressure systems) contain large internal volumes. For example, the PX01 pressure transducer sold by Omega Engineering, Inc., (Stamford, Conn., USA) which is capable of measurement up to 200 megapascals (MPa), has an internal volume of 0.51 milliliter. HPLC solvent delivery systems, which may include multiple pumps sourcing respective mobile phases, typically generate a mixture of two liquids, such as water and an organic solvent in the case of reversed-phase chromatography. In a gradient mode of chromatography, the two liquids are pumped at flow rates that vary over the course of a separation, with the respective flow rates corresponding to programmed profiles. Assume that a separation is performed at a constant flow rate of 4 microliter/minute and that its duration is 5 minutes, over which time the composition of the mobile phase being delivered to the chromatographic column is supposed to vary linearly from 5% organic solvent to 90% organic solvent. This is illustrated schematically in FIG. 1A, where the percentage of organic solvent, measured slightly ahead of the column, is plotted versus time (dash-dotted line 100). At time t=0, the pump controller receives the command to begin a linear ramp from 5% to 90% organic solvent composition. Because of the internal volume of the pump, there is a lag 110, commonly referred to as delay time, before the actual beginning of the gradient. This lag is sometimes also expressed in terms of delay volume, which is simply the product of the delay time and the flow rate. For reasons of speed and performance, this delay time should ideally be as small as possible. If a sensor with a large internal volume is located between the pump and the column, the delay time can be significantly increased, as shown schematically by a larger delay time 125 in FIG. 1A for dashed line 150. For example, if the internal volume is 400 microliters and the flow rate is 4 microliter/minute, the delay time would be increased by 100 minutes which is unacceptable in almost all situations, but especially in a separation that is supposed to last only 5 minutes. In such a separation, the delay time is ideally less than one minute, implying a delay volume of less than 4 microliters in this example.
Furthermore, the fluid volume inside commercially available sensors is often poorly swept. For example, the PX01 pressure transducer has only one fluid access port. When it is used to measure the pressure of a mixture of liquids flowing along a tube or channel and when the composition of this mixture changes over time, the liquids flowing along the tube or channel will mix slowly with the liquid contained inside the sensor, which will change the composition of the mixture. This is undesirable in gradient chromatography, where fidelity to a given composition is highly desired. Ideally, as is illustrated in FIG. 1B by the dash-dotted line 170, corners 175 of the linear ramp of a gradient of composition are sharp. If a sensor with a large internal volume and poor sweeping of this internal volume is placed between the pump and the column, the profile of the gradient will be more like dashed line 180 with highly rounded corners 185.